Geoscience Reference
In-Depth Information
Fig. 12.2 Abitibi area, Canadian Shield; study area as outlined in Fig.
4.12
.(a) Linear model used
to correlate volcanogenic massive sulphide deposits with same lithological variables as used for
40 km) cells are based on sum of estimated posterior probabilities
summed for the 16 (10
10 km) cells contained within these larger unit cells multiplied by the
constant F assuming zero mineral potential within “control” area consisting of all (10
10 km)
“control” cells containing one or more deposits plus all cells that have one or more sides in
common with a control cell. (b) Jackknife estimate as obtained without control area (modified
from Agterberg
1973
, Fig. 4)
Y
j
Y
¼
s
2
s
2
35 control cells each containing
one or more volcanogenic massive sulphide deposits. These were randomly divided
into
k
=
k
. In this example, there were
N
¼
5 control cells. A stepwise solution similar to the one
previously shown in Table 4.4 was obtained for the entire data set (35 control cells)
after setting observed values of
Y
equal to 1 in control cells (Fig.
12.2a
). The contours
in this diagram are for expected number of (10
¼
7 groups of
n
¼
30 km) unit
cell. For this experiment, cell composition data for 8 different rock types in 1,086
(10
10 km) cells per (30
10 km) cells were used as explanatory variables augmented by the 28 cross-
product variables for these eight rock types. A control area surrounding the relatively
well explored mining districts was defined to correct for bias due to undiscovered
deposits. Next, similar stepwise solutions were obtained for 7 additional runs, each
based on 30 control cells. A control area surrounding the relatively well explored
mining districts was not defined for these seven new experiments.
An example of results obtained by this jackknife experiment on Abitibi
volcanogenic massive sulphides is as follows (
cf
. Agterberg
1973
, p. 8).
The explanatory variable first selected when all 35 control cells are used is the
Search WWH ::
Custom Search